Meta Platforms has created smaller versions of its Llama artificial intelligence models that can run on smartphones and tablets, opening new possibilities for AI beyond data centers.

The company announced compressed versions of its Llama 3.2 1B and 3B models today that run up to four times faster while using less than half the memory of earlier versions. These smaller models perform nearly as well as their larger counterparts, according to Meta’s testing.

The advancement uses a compression technique called quantization, which simplifies the mathematical calculations that power AI models. Meta combined two methods: Quantization-Aware Training with LoRA adaptors (QLoRA) to maintain accuracy, and SpinQuant to improve portability.

This technical achievement solves a key problem: running advanced AI without massive computing power. Until now, sophisticated AI models required data centers and specialized hardware.

Tests on OnePlus 12 Android phones showed the compressed models were 56% smaller and used 41% less memory while processing text more than twice as fast. The models can handle texts up to 8,000 characters, enough for most mobile apps.

Tech giants race to define AI’s mobile future

Meta’s release intensifies a strategic battle among tech giants to control how AI runs on mobile devices. While Google and Apple take careful, controlled approaches to mobile AI — keeping it tightly integrated with their operating systems — Meta’s strategy is markedly different.

By open-sourcing these compressed models and partnering with chip makers Qualcomm and MediaTek, Meta bypasses traditional platform gatekeepers. Developers can build AI applications without waiting for Google’s Android updates or Apple’s iOS features. This move echoes the early days of mobile apps, when open platforms dramatically accelerated innovation.

The partnerships with Qualcomm and MediaTek are particularly significant. These companies power most of the world’s Android phones, including devices in emerging markets where Meta sees growth potential. By optimizing its models for these widely-used processors, Meta ensures its AI can run efficiently on phones across different price points — not just premium devices.

The decision to distribute through both Meta’s Llama website and Hugging Face, the increasingly influential AI model hub, shows Meta’s commitment to reaching developers where they already work. This dual distribution strategy could help Meta’s compressed models become the de facto standard for mobile AI development, much as TensorFlow and PyTorch became standards for machine learning.

The future of AI in your pocket

Meta’s announcement today points to a larger shift in artificial intelligence: the move from centralized to personal computing. While cloud-based AI will continue to handle complex tasks, these new models suggest a future where phones can process sensitive information privately and quickly.

The timing is significant. Tech companies face mounting pressure over data collection and AI transparency. Meta’s approach — making these tools open and running them directly on phones — addresses both concerns. Your phone, not a distant server, could soon handle tasks like document summarization, text analysis, and creative writing.

This mirrors other pivotal shifts in computing. Just as processing power moved from mainframes to personal computers, and computing moved from desktops to smartphones, AI appears ready for its own transition to personal devices. Meta’s bet is that developers will embrace this change, creating applications that blend the convenience of mobile apps with the intelligence of AI.

Success isn’t guaranteed. These models still need powerful phones to run well. Developers must weigh the benefits of privacy against the raw power of cloud computing. And Meta’s competitors, particularly Apple and Google, have their own visions for AI’s future on phones.

But one thing is clear: AI is breaking free from the data center, one phone at a time.

Decentralized social app Bluesky announced on Thursday that it has raised a $15 million Series A round, following its $8 million seed raise last year. This funding comes as Bluesky sees increased growth, in part from X users who are troubled by recent changes to the block feature, as well as the move to allow third parties to train AI on users’ public posts. Within the last month alone, Bluesky has added around 3 million new users, bringing its total user base to about 13 million.

Bluesky was initially incubated inside Twitter as former CEO Jack Dorsey’s vision for what the future of social media should look like. But the social network and developer of the open source AT Protocol is no longer affiliated with Dorsey, who left the startup’s board earlier this year. Still, many of the initial goals for Bluesky remain consistent: like Mastodon, Bluesky’s AT Protocol is decentralized, meaning that individual people will be able to set up their own social servers and apps, and people outside of the company have transparency into how and what is being developed.

“With this fundraise, we will continue supporting and growing Bluesky’s community, investing in Trust and Safety, and supporting the ATmosphere developer ecosystem,” Bluesky’s blog announcement reads. “In addition, we will begin developing a subscription model for features like higher quality video uploads or profile customizations like colors and avatar frames.”

The Bluesky team has been quick to tell users that this paid tier will not be like X, where subscribers get exclusive blue check marks and algorithmic up-ranking, making their posts more visible.

“The way twitter did subscriptions was basically a blueprint for how bluesky shouldn’t do them,” Bluesky developer Paul Frazee posted. “‘Pay to win’ features like getting visibility or having a bluecheck because youre a subscriber is just wrong, and ruins the network for everyone.”

The Series A round is led by Blockchain Capital with participation from Alumni Ventures, True Ventures, SevenX, Darkmode’s Amir Shevat, and Kubernetes co-creator Joe Beda. The presence of a crypto-focused firm might alarm skeptics, especially since CEO Jay Graber used to be a software engineer for a crypto company, Zcash, but Bluesky has proactively assured users that the company is not pivoting to web3.

“Our lead, Blockchain Capital, shares our philosophy that technology should serve the user, not the reverse — the technology being used should never come at the expense of the user experience,” Bluesky said in its announcement. “This does not change the fact that the Bluesky app and the AT Protocol do not use blockchains or cryptocurrency, and we will not hyperfinancialize the social experience (through tokens, crypto trading, NFTs, etc.)”

Graber also announced that Kinjal Shah, a General Partner at Blockchain Capital, will be joining the board of Bluesky.

“[Shah] shares our vision for a social media ecosystem that empowers users and supports developer freedom, and it’s been a great experience working with her. With her support, we are well positioned to grow,” Graber wrote.

Bluesky is working on a premium subscription that will add features like higher-quality video uploads and some profile customization options. Unlike the premium subscription offered by X, however, Bluesky’s paid tier won’t boost the visibility of your posts, nor will it give your account a “verified” status. Bluesky, in a post on its blog, also notes that Bluesky “will always be free to use.”

“Subscription revenue helps us improve the app, grow the developer ecosystem, and gives us time to explore business models beyond traditional ads,” Bluesky chief operating officer Rose Wang wrote in a post. “Paid subscribers won’t get special treatment elsewhere in the app, like upranking premium accounts or blue checks next to their names.”

The iKOOLCORE R2 Max is a compact yet powerful mini PC that comes with either the Intel N100 or the more powerful Intel Core i3-N305, making it capable of handling various tasks such as content creation, virtualization, and office work.

Despite its small size, measuring just 15.7 x 11.8 x 4 cm, the R2 Max is well-equipped with four high-speed Ethernet ports – two 10GbE ports powered by Marvell AQC113C-B1-C chips and two 2.5GbE ports running on Intel i226-v controllers.

Tech billionaire Jared Isaacman became the first civilian to conduct the spacewalk on Thursday. Teaming up with SpaceX, the five-day Polaris Dawn mission was launched into space with a four-member crew on Tuesday from NASA’s Kennedy Space Center in Florida. 

Joining the small elite group of spacewalkers, Isaacman and his crew waited until the capsule was depressurised before opening the hatch. 

ALSO READ | Polaris Dawn mission: Blast off! SpaceX launches all-civilian crew for first privately-funded spacewalk

All the crew onboard wore SpaceX’s new spacewalking suits, as the entire capsule was exposed to the vacuum of space. The suits offered protection to all the crew members from the harsh vacuum. 

“From here Earth sure looks like a perfect world,” said Isaacman after the spacewalk.

The orbit was reduced by half to 458 miles (737 kilometres) for the spacewalk.

ALSO READ | Human health in space: Charting the unknown on Polaris Dawn mission

Testing the spacewalking suit was one of the important tasks for the team. The spacewalking was expected to last for about two hours. Rather than the exact walking, it mostly involved stretching and movement of limbs. 

Isaacman had his hand or foot always attached to the capsule as he flexed his arms and legs. The hatch sported a walker-like structure for extra support. These are considered EVA- extravehicular activities, any activity done by an astronaut in outer space outside a spacecraft. 

According to experts, due to the microgravity situation nobody really ‘walks’ on the space. Any activity outside the spacecraft can be considered as spacewalking as there is movement outside the capsule in space, experts pointed out. 

After about 15 minutes outside, Isaacman was to be replaced by the SpaceX engineer Sarah Gillis to go through the same motions. 

So far about 263 people have conducted a spacewalk representing 12 countries. 

When pushed into the tiniest of cracks, water can be used in unexpected ways. A new battery design that relies on tiny amounts of water confined within layers of clay could eventually offer sustainable power in places as extreme as Mars.

Vasily Artemov at the Swiss Federal Institute of Technology and his colleagues built the battery with components similar to conventional batteries, including two electrodes, one with a negative and one with a positive charge. But instead of making these electrodes metal, they used the carbon-based material graphene. Instead…